Text Identification for Questions Generation According to Bloom's Taxonomy Using Natural Language Processing

Adaptivity is actively pursued for testing in education as it provides alternative paths of action for the students. Particularly in engineering education adaptive assessments provide an estimation of students’ competence to provide novel solutions and accurate evaluation of knowledge. This chapter discusses the question identification based on Bloom’s taxonomy for the design of Computerized adaptive assessment. Indicators selected for classification are pre-determined verbs that represent the Bloom’s stages indicating the level of course expertise. Questions are labelled as per Blooms taxonomy. Natural Language Processing is a Machine learning based method to classify the text, as it easily converts the text into tokenized format that can be used for classification. Classification models are developed using the Bidirectional Long Short-Term Memory networks and Convolutional Neural Network. Deep Learning toolbox of MATLAB 2023a is used to train and test the model. The CNN model has provided a validation accuracy of 85.71% for the labeled questions.